.insert_df fails to insert df column in UInt256 ch column

[Dynortice]

Describe the bug

When attempting to use .insert_df to insert a DataFrame into a ClickHouse table that includes a UInt256 column, the operation fails with an AttributeError. Specifically, it appears that the code attempts to call the to_bytes method on a numpy.int64 object, which does not have this method even when dtype of column is Python built-in int.

Steps to reproduce

1. Create a ClickHouse table with a UInt256 column.
2. Use the insert_df method a pandas DataFrame containing a column meant to populate the UInt256 column.
3. Observe error.

Expected behaviour

The .insert_df method should successfully convert the pandas DataFrame column to the appropriate format for insertion into the UInt256 column in ClickHouse.

Code example

import clickhouse_connect
client = clickhouse_connect.get_client(host='localhost', port=9000, username='default', password='')
client.command('CREATE TABLE test_table (id UInt256) ENGINE = Memory')
data = pd.DataFrame({'id': [1, 2, 3]})
client.insert_df('test_table', data)

clickhouse-connect and/or ClickHouse server logs

Traceback (most recent call last):
  ...
AttributeError: 'numpy.int64' object has no attribute 'to_bytes'. Did you mean: 'tobytes'?

Configuration

Environment

• clickhouse-connect version: 0.8.13
• pandas version: 2.2.3
• Python version: 3.12.0
• Operating system: Windows 11 Pro 22631.4602

ClickHouse server

• ClickHouse Server version: 23.6.2.18